521–530 of 609 posts
I echo what others say, Kagi is a joy to use and feels just like Google used to be - useful
I use DuckDuckGo, with the occasional reddit !g appended if I'm looking for something experience-based. For me, searches fall into one of three categories, none of which are a good fit for LLMs: 1. A single business, location, object, or concept (I really just want the Google Maps or Wikipedia page, and I'm too lazy to go straight to the site). For these queries, LLMs are either overkill or outdated. 2. Product revie…
I've tried DDG but it doesn't seem much different from the results on Google/Bing (as I understand it they use Bing's search index anyway?)
For more general searches, depending on the topic, DDG is close to useless because link farms, AI slop, and returning results that aren't really what I'm looking for (some of the keywords weight too much). But I suspect this is a common problem in all search engines, so I'm not looking for a replacement. It is frustrating though. I can't believe the information doesn't exist, is just that it is unreachable.
I don't search using AI. Generally I'm not looking for information that can be distilled into an "answer"; and there's also that DDG is not feeding me AI answer (I think? May be I'm not paying attention).
1. Bookmark manager. I can write "maven download sources", click on baeldung and copy&paste command from there. I did that 100 times and I'll do it 101-th time. I have plenty of webpages that I know they exist and I know how to find them. I'm too lazy to actually bookmark and organize them and Google works just fine for me.
2. Search for new knowledge in general domains. This category of queries I sometimes use ChatGPT for it, but not always. It's hard to formulate the rules, but I have a feeling which tool is more useful for given question. Sometimes it's just quicker to type few keywords in Google rather than asking full-fledged question to ChatGPT and wait for it to return full-fledged answer.
3. Search for new knowledge in arcane domains. For example I might need new info about my city, where to buy laptop, etc. ChatGPT might know about US, but its knowledge about Kazakhstan is definitely limited, shallow and outdated, so real web comes to rescue here.
4. Using Google-specific functions. For example I very often use queries like "site:bla.com something", because website-hosted searches in 100% cases are braindead and I wouldn't even bother with it.
For me, ChatGPT main function is to create new content. For example I might want to start new Go project, so I'll ask ChatGPT to write hello world golang tutorial for me, then ask follow-up questions like what identifier naming should I follow, etc. There are, of course, hundreds of similar articles in the Internet, but ChatGPT delivers tailored data much faster.
Yes, I do. I'd never use an LLM for any meaningful or important information because by design, they will simply shove the most likely next token and you get a ton of responses which are pure nonsense if you start digging into it. Mind you, I've been noticing that Google has been terrible for a long while now. Kagi seemed alright at first but it also gave a lot of nonsense. The final straw for Kagi was the fact that t…
Google seems to be better at bringing up a variety of stackoverflow and blogposts relevant to my search queries. Qwant seems to struggle exactly with that: it's great at giving exactly what I was searching for but that's sometimes not what I was looking for, if you get what I mean.
In a sense, LLM's are actually perfect for that. But like you say, the super confident hallucinations are just too frustrating. Literally every time I've asked it a serious programming question, it's hallucinated an API that doesn't exist. Everyone seems to be focusing on letting LLM's solve math and thinking problems. That's exactly what _I'm_ good at. I would much rather have an LLM that is good at combining sources and giving me facts (knowledge, rather than thinking) while most of all being able to say "I don't know".
Given my time dedicated to researching thing, I feel like I am "more productive" b/c I waste less time.
But I do my due diligence to double-check what ChatGPT suggests. So if I ask ChatGPT to recommend a list of books, I double-check with Goodreads and Amazon reviews/ratings. Like that. I guess it's like having a pair-research-sesson with an AI librarian friend? I am not sure.
But I know that I am appreciative. Does anyone remember how bad chatbots were before the arrival of low-hanging-AI-fruits like generative AI? Intel remembers.
Search is primarily a portal - you know a particular resource exists, you just don't know its exact URL. You hear about this new programming language called "Frob", and you assume it must have a website. So you google "Frob language". You hear that there was a plane crash in DC, and assume (CNN/AP/your_favorite_news_site) has almost certainly written an article about it. You google "DC plane crash." LLMs aren't ever…
I'd go further, and say I use search when I'm pretty confident I know the right search terms. If I don't, I'll type some wordy long explanation of what I want into an LLM and hope for the best. The reason is pretty simple. If the result you want is in the first few search hits, it's always better. Your query is shorter so there is less typing, the search engine is always faster, the results are far better because you…
Most often not true in the kind of searches I do. Say, I search for how to do something in the Linux terminal (not just the command, but the specific options to achieve a certain thing). Google will often take me to pages that do have the answer, but are full of ads and fluff, and I have to browse through several options until I find the ones I want. ChatGPT just gives me the answer.
And with any halfway decent model, hallucination only seems to be a problem in difficult or very specialized questions. Which I agree shouldn't be asked to LLMs (or not without verifying sources, at least). But over 90% of what I search aren't difficult or specialized questions, they're just things I have forgotten, or things that are easy but I don't know just because they're not in my area of expertise. For example as a learner of Chinese, I often ask it to explain sentences to me (translate the sentence, the individual words, and explain what a given word is doing in the sentence) and for that kind of thing it's basically flawless, there's no reason why it would hallucinate as such questions are trivial for a model having tons of Chinese text in its training set.